The detection of ischemic heart disease has been hampered by problems inherent in conventional imaging, thus leading to: (1) Large overlap in image-derived parameters of function between regions of infarction, ischemia, and nonischemic dysfunction. (2) Poor observed correlation between these parameters and the extent and degree of transmurality of infarction. Additionally, the endocardium is the myocardial layer most affected by ischemia, yet no imaging method can separately assess endocardial thickening. Evidence indicates that regional asynchrony in early systole and early diastole may be a more sensitive indicator of ischemia than end diastolic (ED) to end systolic (ES) comparisons, which are the most commonly used. A major limitation of conventional imaging, thought to be, at least in part, responsible for these deficiencies, is the lack of a practical method (short of implanting physical markers) to spatially and temporally track specific points in the myocardium. The complex rigid body motion, deformation, and temporal heterogeneity exhibited by the heart during its contractile sequence are therefore not accounted for by current imaging approaches. In the absence of such a method, a major unresolved question is to know whether cardiac imaging is fundamentally limited because of intrinsic physiologic overlap between normal and ischemic myocardium or whether better measures can be obtained if the current limitations of cardiac imaging are corrected. We propose to address this issue with a method based on the concept of myocardial tissue tagging by MRI.This method permits dynamic 3-D mapping of the entire left ventricle with 8 to 12 volume elements per slice subdivided into endocardial and epicardial regions, thus allowing a comprehensive analysis of local deformation with a temporal resolution of 25 to 50 ms. We will determine whether this method can provide better indices for detecting and quantifying regional ischemia than conventional analyses limited to ED and ES measurements. The main hypotheses to be tested are that time-dependent indices of regional function derived from dynamic 3-D tagged MRI can (1) more effectively discriminate ischemic from normal myocardium and (2) more effectively correlate with size, location, and extent of ischemia, infarction, or nonischemic dysfunction than conventional methods based on 2-D ED/ES analyses. If proven, these hypotheses would make possible the application of this noninvasive methodology to ischemic heart disease in man.